Research On Grading Technology And Control System Of Citrus Surface Color

The post-harvest technology of Citrus in China is developing continuously.At present,the sorting method of citrus color mainly stays at the level of manual operation,which has the characteristics of strong subjectivity and low efficiency.Therefore,how to improve the accuracy of citrus color grading and production efficiency is the current problem to be solved.In this paper,the method of color grading and defect recognition of Citrus surface is studied by using machine vision detection technology and automatic control technology,and a citrus color grading control system based on machine vision is designed to realize color detection and defect detection of citrus.The main research work and conclusions are as follows:1)The image recognition algorithms for citrus defect and color detection are studied,including image preprocessing,defect and color feature extraction.In the aspect of Citrus image preprocessing,several image preprocessing algorithms are compared and studied,including smoothing and denoising,image segmentation and so on.The experimental results show that the median filtering algorithm is the best for smoothing and denoising,and the OTSU algorithm is the best for image segmentation.In the aspect of color detection,RGB color space is converted into HSV color space to obtain.H component parameters,and the hue mean method and hue interval statistics method were used to detect the color consistency of citrus.The experiment showed that the hue interval statistics method had higher detection accuracy.In defect detection,the method of removing peduncle based on roundness was used to eliminate the influence of peduncle on defect detection,and the ratio of defect area to fruit area was used to realize the defect detection of citrus.2)Based on the requirement of citrus color grading,a citrus color grading control system based on machine vision was designed,including image acquisition system,motion control system and grading execution system.The integration of image acquisition system and control system and mechanical system was completed.The citrus image acquisition and recognition software and grading control software system based on PLC were developed.3)Based on the above-mentioned design of citrus color grading control system,defect detection and color grading of citrus were tested.The experimental results show that the accuracy of defect detection by removing pedicels is 6%higher than that without removing pedicels.The overall accuracy of detection is 92%.The color detection divides citrus into four color grades:orange fruit,yellow fruit,light green fruit and dark green fruit.The overall classification accuracy was 93.75%.